James: > I agree that we need to pay heed to the intent of the author. > > However, if all that we deliver is the .so.1 file without the > compilation symlink and without the header files needed to use the > library, users aren't going to run into this very easily. > > We can always make it Committed in the future if it turns out that > there really is more than just this one consumer.
Indeed. >> Quality manpages and good documentation are probably the sort of thing that >> attracts developers, no? > > Absolutely. We can't do a half-baked job here. It should be either a > good first-class citizen on its own merits, or should be buried > deeply where it won't hurt anyone. > > The project team seems to assert that it belongs in the latter > category. Are you sure it's in the former? It's been around for a > while, but I can't see much that has decided to use it. The web page > says ifplugd, Avahi, ivam2, Nautilus-Share, and aeswepd. > > Of those, only Avahi seems to have any applicability to any future > version of [Open]Solaris. The rest are seem to be duplicates of > existing features/projects. This is up for Padraig to specify. I just was complaining that the JDS team (and probably other teams at Sun) probably needs to take a bit more care when thinking about whether we want/need manpages. Just because we consider something private, doesn't mean we shouldn't ship a manpage when the interfaces are public. This is really a case-by-case issue when dealing with free software. Just my opinion. Brian
